Journal of Intelligent & Fuzzy Systems 18 (2007) 99108 IOS Press

99

Fuzzy SWOT analysis

Abstract. The SWOT (which is sometimes called TOWS) matrix is one of the most important tools for strategic planning specially in the stage of extracting strategies. While the use of SWOT is quite common and popular, it still continues to have certain structural problems. The most important of which are the lack of considering uncertain and two sided factors, lack of prioritization of the factors and strategies and too many extractable strategies. This paper attempts to solve some of the problems by following the fuzzy approach to the internal and external factors (in the form of fuzzy membership functions). The presented algorithm in this article prioritizes and extracts the most signicant strategies based on the intensity of the effect. Keywords: Fuzzy, SWOT, strategic planning

1. Introduction SWOT (an acronym standing for Strengths, Weaknesses, Opportunities and Threats) analysis is a commonly used tool for analyzing internal and external environments in order to attain a systematic approach and support for decision making. The SWOT approach is based on the aggregation of the internal (strengths, weaknesses) and external (opportunities, threats) factors for adopting strategies. In other words, the extracted strategies of SWOT matrix is comprised of four categories of factors combinations: Strengths and Opportunities (S-O); Strengths and Threats (S-T); Weaknesses and Threats (W-T); Weaknesses and Opportunities (W-O).

competency-based planning, however [2] could point out the relation between SWOT and two above approaches. Despite its wide applications, the SWOT method has also a number of problems that 7 of them are mentioned by Hill & Westbrook [3], but most important ones are as follows: I- Usually only qualitative examination of environmental factors is considered [1]; II- It considers no priority for various factors and strategies; III- If the number of factors are more, the number of adopted strategies will be increased exponentially (for example if the number of each set of factors of S, W, O, T is equal to 5, the resulting number of the combined strategies will be around 100 which would make the selection of the appropriate strategy very difcult); IV- It does not consider the vagueness of the factors. In most situations, it is not possible to differentiate the factors clearly either they are opportunities or threats. For example, having no powerful competitor in the market is an opportunity (for controlling the market), and at the same time it may be a threat too (because it may lead the company to inaction and depression). As another example, governmental supports of the company may be both an opportunity (for utilizing governmental facilities) and a threat (since it may cause

Although the SWOT analysis date backs to 60s [9], Weihrichs article [17] that introduces SWOT matrix as a tool for situation analysis, may be the most important references in this eld that has provided some classic examples. After that time, SWOT has been pointed in the most of references of strategic planning. Of course this analysis seems outdate in comparison with recent approaches such as resource-based planning and
Corresponding author. [email protected].

Tel./Fax: +98 2166903067; E-mail:

1064-1246/07/$17.00 2007 IOS Press and the authors. All rights reserved

100

S. Ghazinoory et al. / Fuzzy SWOT analysis

many problems for the company whenever these supports are cut). As another example, economic ourishing is both an opportunity (since it raises the demand for the products of the company) and also a threat (since it would cause a rise in the prices of the raw materials and machineries needed for the company). 1 Therefore in such ambiguous cases the use of fuzzy sets is justied to be applied. In fact a factor with certain membership value belongs to one of the categories. Hence we recommend the use of fuzzy SWOT matrix. For example economic ourishing is an opportunity with 0.7 as membership value and it is a threat with membership value 0.3. The purpose of this article is to use of this concept for combination of internal and external factors so that, SWOT analysis can considers uncertainty of factors and determines the priority of strategies. It seems this method helps to increase the effectiveness of SWOT methodology and extends its applications.

2. Application of the fuzzy theory in the strategic planning Although up to now, efforts have been made to solve out the SWOT problems, the usual approach is the use of prioritization methods. In particular, the Analytic Hierarchy Process (AHP) has been implemented for this case [57]. Kurttila et al. [6] developed a hybrid method in order to improve the usage of SWOT analysis and to eliminate the deciencies about measurement and evaluation [3,12] by a systematical approach. The method, which is called AWOT in [4,10], is in fact obtained by connecting the Analytic Hierarchy Process (AHP) with SWOT analysis. In the studies mentioned above, similar to [6], only weights of the SWOT groups and factors are determined. In the hierarchical structures based on the strategic factors, the strategies or the alternatives did not take part. Obviously in usual methods of multi criteria decision making such as AHP [15] and also fuzzy AHP [16], the aim is to sort the strategies. On the other hand we aim a methodology to form, evaluate and prioritize the
1 Of course in some cases, one factor is listed in both opportunities and threats. For example internet for campus universities in [2], but it is an exception because if we do this for most of factors, the number of factors will extremely increase and previous problem will be intensied.

strategies and since the information in this respect is not precise, therefore employing fuzzy sets is justied. In SWOT analysis some uncertainties can be encountered in the evaluation of the organizations internal and external environment. It can be stated that it is impossible for the opportunities and threats arising from the external environment of the rm to be always denite in any condition. Consequently, it is also impossible to measure the numerical values precisely. Similarly, it is not always possible to measure or evaluate the strengths and weaknesses which are the consequences of the assets and abilities of the rm exactly. There is a similar case in the SWOT factors. For example, it is not easy to measure the exact values for political uncertainties or possible problems in the environment, both strategic factors and threats for the organization. Besides, in SWOT analysis, it is difcult to evaluate the factors in any case and at anytime in a binary based Aristotle logic. For instance, it may not be realistic to represent the image about the goods and services produced by the rm in two points as sufcient (1)/ insufcient (0). In real life, there may be different evaluations between these two. Therefore, it will not be possible to understand the actual situation with an evaluation which does not assume fuzziness. The application of fuzzy concept in interpretation of portfolio matrices has been recommended by [11] and [14]. In [11], the Industry Attractiveness-Business Strength matrix is fuzzied using fuzzy weighted average [14] employs a fuzzy rule-base to handle the Growth-Share matrix. Despite these researches and because of different structure of SWOT matrix, we believe that different approach should be applied for the SWOT matrix. However, the notion of fuzzy theory has been applied in strategic planning, but to our information, it seems that there is a high scope for application of fuzzy sets in SWOT matrix. In todays rapidly changing and highly uncertain environment, the strategic decisions have an extremely complex and fuzzy nature [1]. In usual approach of SWOT analysis, strategists have tendency to consider importance of factors and strategies as the most signicant criterion to select the most appropriate strategies but it was not possible to do it systematically up to now. In what follows, we propose an integrated approach to consider the ambiguity of internal and external factors as the inputs of strategic planning process and to make an environment to prioritize the factors and strategies for selecting the most appropriate alternatives.

S. Ghazinoory et al. / Fuzzy SWOT analysis

101

-10

Weakness

0
Fig. 1. A scale representation.

Strength

+10

3. The algorithm for fuzzifying the SWOT matrix In this section an algorithm is presented for rectifying the shortcomings and problems of the SWOT matrix through the use of fuzzy sets. The steps taken for this algorithm are as follows: 3.1. Scaling the factors For each internal factor a suitable membership function is developed in the range 10 to 10. As each factor may exhibit two sided meaning, therefore the negative part shows the weakness and the positive part of the fuzzy membership function signies the strength of the factor. Figure 1 for instance, illustrates such scaling with a Gaussian membership function. The same statement is true for external factors too. The reason of using membership function can be understood using Fig. 1. In dened domain between the most intensive weakness (10) and the most intensive strengths (+10), each factor can specify the level of its belonging to each one of domain points and the points that its major concentration is on, using proper fuzzy membership function. In dened domain between the most intensive weakness (10) and the most intensive strengths (+10), each factor can specify the level of its belonging to each one of domain points and the points that its major concentration is on, using proper fuzzy membership function. The type of membership function is optional but to simplify the computation of the area location in next stages, we suggest the triangular membership functions. A triangular membership functions can be specied by three parameters (x p , xm , xo ) as follows [13]:

y = triangle(x; xp , xm , xo ) 0, (x xp )/(xm xp ), = (xo x)/(xo xm ), 0,

x xp xp x xm x xo x

xm x0

Where in this paper, x p , xm and xo are called pessimistic, probable and optimistic values respectively. We can make a list of factors including strengths, weaknesses, opportunities and threats. The difference between this stage and the usual SWOT is that for this case, 3 questions should be asked for each internal factor from the decision maker: What value to be allocated in the range 10 to +10 for pessimistic (xp ), probable (xm ) and optimistic (xo ) situations for internal factor (10 to 0 for the intensity of the weakness and 0 to +10 for the intensity of the strength)? The same spread of 10 to 10 is also assumed for the external factors by the same way of questions and a triangular membership function would be obtained for each factor. 3.2. Aggregation of membership functions of internal and external factors To extract strategies on the basis of internal and external factors, it is necessary to aggregate the membership functions of the factors. In this regard, we dene a three dimensional surface based on the membership functions of internal and external factors. This is applicable using a suitable t-norm such as min operator. Each point of this surface in SWOT matrix is derived from the intersection of membership functions of relative external and internal factors i.e.:

102

S. Ghazinoory et al. / Fuzzy SWOT analysis

Fig. 2. An aggregated graph of internal and external membership functions.

Fig. 3. Surface created by different (2 > 1 ).

S (x, y) = min{I (x), E (y)} Where S (x, y) is the membership function of fuzzy SWOT matrix and I (x) and E (y) are the member-

ship function for internal and external factors respectively. A sample of such 3D surface is presented in Fig. 2.

S. Ghazinoory et al. / Fuzzy SWOT analysis

103

Internal factor axis (x)

-10

+10

+10

70 %

10 % External factor axis (y)

0 20 %

-10
Fig. 4. Top view of resultant surface by -cut.

Extraction of strategies necessitates the decrease in vagueness within membership functions [11]. To do this, we dene a specic -cut on the surface, produced from previous stage. The value of depends upon the level of uncertainty we are considering. Increasing , decrease the surface over the -cut. Thus, we will be encountered with less uncertainty and vice versa. Usually the strategists prefer to choose around 0.5. Because the value of close to zero increases the amount of uncertainty to large extent so that deciding on proper strategy creates a complex problem. On the other hand, tending to 1 gives a less domain for interaction. Figure 3 demonstrates the surface created by different . Figure 4 illustrates the section of the surface (area) created by -cut corresponding Fig. 3 ( 1 ). Now by projecting this surface onto the SWOT matrix further evaluations can be made.

3.3. Evaluation, prioritization and extracting strategies The basis for evaluation of the areas resulting from previous stages is their distance from the corners of the matrix. Different criteria may be dened for this distance. 1) Min of the Euclidean distance of each corner point from the boundary of the area, 2) min of the distance of each corner point from the center of gravity of area, 3) max of the percentage of area in each quadrant. Selection of each of these 3 criteria will not leave an impression on the prioritizing, but introducing them is necessary for better understanding the concept of distance. In this paper, the areas which are near the corners of matrix are preferred to extract the strategies. It is necessary to consider the criteria for evaluation and prioritization, but it is not sufcient since if the factors are not correlated, it will not make any sense to extract the strategies.

104

S. Ghazinoory et al. / Fuzzy SWOT analysis Table 1 Internal factors evaluation Row a b c d e f g h i j k Internal factors High inuence of the share holders of the company in the government Prot of company has increased in last year New products have succeeded Market share has increased in Europe Company has received HACCP certicate The possibility of gaining large amount of loans from the bank Low potentials of the experts and engineers Sale of one old product has decreased Cost of maintenance for old facilities and machines has increased Weakness in marketing Low present prot in proportion to standard level Table 2 External factors evaluation Row q r s t u v w x y z External factors Development of economical systems based on free market in south east Asia Demand has increased for products of company in foreign market High sensitivity for safeness of food stuffs in foreign market Duty free laws in some countries Prohibition against the imports of the foreign competitive products Rapid growth of the market in recent periods Cost of production is increasing with more acceleration in proportion to income of company Low share of market comparing other companies Uncertainty of consumption market in south east Asia Disliking of some customers to consume ready foods in south east Asia Value (2, 0, 2) (1, 1, 2) (2, 1, 1) (1, 0, 3) (2, 1, 2) (0, 3, 4) (3, 2, 0) (2, 1, 1) (3, 1, 1) (2, 1, 0) Value (1, 2, 3) (1, 2, 3) (2, 1, 3) (2, 1, 3) (0, 2, 3) (2, 1, 3) (5, 4, 1) (3, 2, 0) (5, 4, 3) (2, 1, 0) (2, 1, 0)

4. A case illustration 4.1. Determining the membership functions for factors In order to evaluate the applicability of the proposed algorithm, we implemented it in a strategic planning process for a food corporation in Iran. The corporation is a successful organization in food industry in Middle East. The company identies important internal factors by questioning its experts with their pessimistic (x p ), probable (xm ) and optimistic (xo ) values as Table 1. In addition, the company would identify its major external factors with pessimistic, probable and optimistic values as Table 2. 4.2. Aggregation of membership functions of internal and external factors Aggregation of membership functions of factors a and q is based on the Fig. 5. Aggregation of the other internal and external factors will be done in the same way. It means that 2 by 2 combinations will be done for all factors and extracted strategies will be prioritized in the next stage. As we can see in Fig. 5, the result of aggregation of these two membership functions will lead to the formation of a pyramid.

Next we should dene a specic value of -cut (say = 0.2) as become closer to 1, uncertainty decrease but exibility of extracted strategies will also decrease instead. By replacing Z with 0.2 in the equations of each side of pyramid, the location of the area could be calculated as shown in Fig. 6. The resultant area of aggregation of factors a and q could be seen in Fig. 7. 4.3. Evaluation, prioritization and extracting strategies To assess the vicinity of areas to corners, we calculate the percentage of areas in each quadrant of SWOT matrix which is presented in Table 3. As we can see in Table 3, beside other factors, the achieved areas for the factors (g/v), (g/w), (g/z), (h/v), (h/w), (h/z), (i/v), (i/w), (i/z), (j/v), (j/w), (j/z), (k/v), (k/w), (k/z) are closer to the corners of matrix. Therefore these strategies are recommended: The resulting strategy from the factors g/v, g/w and g/z may be training the engineers to reduce the cost of production and training the commercial experts for powerful marketing and advertising in south East Asia.

S. Ghazinoory et al. / Fuzzy SWOT analysis

105

Fig. 5. Aggregation of membership functions of factors a and q.

Fig. 6. -cut with = 0.2.

The resulting strategy from the factors h/v, h/w, and h/z may be to remove the old products for reducing cost of production and opposing customer dissatisfaction. The resulting strategy from the factors i/v, i/w, and i/z may be replacement of old equipment to respond to the needs of the market and price reduction. The resulting strategy from the factors j/v, j/w, and j/z may be investment in marketing to improve the market share and customer satisfaction.

It is noteworthy to mention that it is not necessary for all 110 areas obtained to be meaningful regions for making strategy. That is, it is possible that some of the internal and external factors may have not relations and therefore we fail to extract a strategy (for example factors k/v, k/w or k/z). In this way, the company may continue to extract strategies from the next priorities as far as it pleases. In classical SWOT analysis for this case, there are 110 alternatives with equal chance to select and it is unclear for strategist which factors should be selected for

106
-10

S. Ghazinoory et al. / Fuzzy SWOT analysis

+10

2 1
External factor axis (y)
Factors a and q

6.25%

43.75 %

6.25% 43.75 %

3 4

-10

Fig. 7. Resultant area of factors a and q.

extracting the most important strategies, while in presented method we know that we should begin with factors which take priority over the others and confusion of deciding in classical method is changed to systematic evaluation and selection of the factors and extraction of strategies, moreover the ambiguity and uncertainty of the factors (which described in Section 1) can be considered during the process of decision making.

Screening strategies and select important ones from a large number; Vagueness of factors is considered using fuzzy theory. In spite of these advantages, this method has its limitations: Fuzzication have different methods and techniques [18] that selecting each of them will leave an impression on results and particularly, value of -cut can make difference in results and fuzzication method and amount of is depended n research and his/her experience. Amount of necessary calculations is almost high like most of other fuzzy techniques. Selecting pair of factors that lead to meaningful strategies is difcult and need experience. Totally, after studying weaknesses and strengths of this method, we can call it a step forward in SWOT analysis.

5. Discussion It was mentioned that although SWOT analysis is widely used in strategic researches, but sill there are limitations that in case of removing them, this analysis will be more efcient and with higher validity. The method that was presented in this article can remove the limitations discussed in introduced to some extent: Fuzzy theory can convert quantitative to membership function in addition to qualitative factors and has recognized methods for this propose [14]; Prioritization of extracted strategies is possible;

S. Ghazinoory et al. / Fuzzy SWOT analysis

107

Table 3 Percentage of areas in each quadrant Factors Percentage of of area in quadrant 1 43.75 65.625 21.875 65.625 49.21875 87.5 0 21.875 16.40625 0 43.75 65.625 21.875 65.625 49.21875 87.5 0 21.875 16.40625 0 32.5 48.75 16.25 48.75 36.5625 65 0 16.25 12.1875 0 32.5 48.75 16.25 48.75 36.5625 65 0 16.25 12.1875 0 35 52.5 17.5 52.5 39.3 70 0 17.5 13.1 0 32.5 48.75 16.25 48.75 36.5625 Percentage of area in quadrant 2 6.25 9.375 3.125 9.375 7.03125 12.5 0 3.125 0 0 6.25 9.375 3.125 9.375 7.03125 12.5 0 3.125 0 0 17.5 26.25 8.75 26.25 19.6875 35 0 8.75 0 0 17.5 26.25 8.75 26.25 19.6875 35 0 8.75 0 0 15 22.5 7.5 22.5 16.8 30 0 7.5 5.6 0 17.5 26.25 8.75 26.25 19.6875 Percentage of area in quadrant 3 6.25 3.125 9.375 3.125 5.46875 0 12.5 9.375 10.15625 12.5 6.25 3.125 9.375 3.125 5.46875 0 12.5 9.375 10.15625 12.5 17.5 8.75 26.25 8.75 15.3125 0 35 26.25 28.4375 35 17.5 8.75 26.25 8.75 15.3125 0 35 26.25 28.4375 35 15 7.5 22.5 7.5 13.1 0 30 22.5 24.3 30 17.5 8.75 26.25 8.75 15.3125 Percentage of area in quadrant 4 43.75 21.875 65.625 21.875 38.28125 0 87.5 65.625 71.09375 87.5 43.75 21.875 65.625 21.875 38.28125 0 87.5 65.625 71.09375 87.5 32.5 16.25 48.75 16.25 28.4375 0 65 48.75 52.8125 65 32.5 16.25 48.75 16.25 28.4375 0 65 48.75 52.8125 65 35 17.5 52.5 17.5 30.6 0 70 52.5 56.8 70 32.5 16.25 48.75 16.25 28.4375 Factors Percentage of area in quadrant 1 65 0 16.25 12.1875 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Percentage of area in quadrant 2 35 0 8.75 0 0 50 75 25 75 56.25 100 0 25 18.75 0 50 75 25 75 56.25 100 0 25 18.75 0 50 75 25 75 56.25 100 0 25 18.75 0 50 75 25 75 56.25 100 0 25 18.75 0 50 75 25 75 56.25 100 0 25 18.75 0 Percentage of area in quadrant 3 0 35 26.25 28.4375 35 50 25 75 25 43.75 0 100 75 81.25 100 50 25 75 25 43.75 0 100 75 81.25 100 50 25 75 25 43.75 0 100 75 81.25 100 50 25 75 25 43.75 0 100 75 81.25 100 50 25 75 25 43.75 0 100 75 81.25 100 Percentage of area in quadrant 4 0 65 48.75 52.8125 65 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

a and q a and r a and s a and t a and u a and v a and w a and x a and y a and z b and q b and r b and s b and t b and u b and v b and w b and x b and y b and z c and q c and r c and s c and t c and u c and v c and w c and x c and y c and z d and q d and r d and s d and t d and u d and v d and w d and x d and y d and z e and q e and r e and s e and t e and u e and v e and w e and x e and y e and z f and q f and r f and s f and t f and u

f and v f and w f and x f and y f and z g and q g and r g and s g and t g and u g and v g and w g and x g and y g and z h and q h and r h and s h and t h and u h and v h and w h and x h and y h and z i and q i and r i and s i and t i and u i and v i and w i and x i and y i and z j and q j and r j and s j and t j and u j and v j and w j and x j and y j and z k and q k and r k and s k and t k and u k and v k and w k and x k and y k and z

108

S. Ghazinoory et al. / Fuzzy SWOT analysis [2] R.G. Dyson, Strategic development and SWOT analysis at the University of Warwick, European Journal Of Operational Research 152 (2004), 631640. T. Hill and R. Westbrook, SWOT Analysis: Its Time for a Product Recall, Long Range Planning 30(1) (1997), 4652. M. Kajanus, J. Kangas and M. Kurttila, The use of value focused thinking and the AWOT hybrid method in tourism management, Tourism Management 25 (2004), 499506. M. Kurttila, J. Kangas, M. Pesonen, M. Kajanus and P. Heinonen, Using AHP and SWOT Analysis in Assessing Priorities of Alternative Strategies in Forest Planning, International Symposium on Advanced Technology in Environmental and Natural Resources, Rovaniemi, Finland, 812 June 1998. M. Kurttila, M. Pesonen, J. Kangas and M. Kajanus, Utilizing the analytic hierarchy process (AHP) in SWOT analysis A hybrid method and its application to a forest certication case, Forest Policy and Economics 1(1) (2000), 4152. M. Kurttila, J. Kangas and M. Kajanus, The use of value focused thinking and the AWOT hybrid method in tourism management, Tourism Management 25(4) (2004), 499506. Y. Lai and C. Hwang, Fuzzy mathematical programming, Springer Verlag, Berlin, 1992. E.P. Learned, C.R. Christensen, K.E. Andrews and W.D. Guth, Business Policy: Text and Cases, Irwin, Homewood, IL, 1965. L.A. Leskinen, P. Leskinen, M. Kurttila, J. Kangas and M. Kajanus, Adapting modern strategic decision support tools in the participatory strategy process-a case study of a forest research station, Forest Policy and Economics, Article (2004), in press. C. Lin and P.J. Hsieh, A fuzzy decision support system for strategic portfolio management, Decision Support Systems 38 (2003), 383398. M.H.B. McDonald, The Marketing Planner, Oxford:Butterworth-Heinemann, 1993. P. Melin and O. Castillo, Modeling, simulation and control of non-linear dynamical systems, Taylor and Francis 11 (2002), New Fetter Lane, London. E. Pap, Z. Bosnjak and S. Bosnjak, Application of fuzzy sets with different t-norms in the interpretation of portfolio matrices in strategic management, Fuzzy Sets and Systems 114 (2000), 123131. T.L. Saaty, The analytic hierarchic process, McGraw-Hill International, New York, 1980. E. Triantaphyllou, Multi-criteria decision making methods: A comparative study, Kluwer Academic Publishers, London, 2000. H. Weihrich, The TOWS matrix-a tool for situational analysis, J. Long Range Plan 15(2) (1982). H.J. Zimmermann, Theory of fuzzy sets and its application, Kluwer Academic Publishers, Boston, 1996.

6. Conclusion In this article some of the weaknesses of the SWOT matrix are pointed out and efforts were made to solve them using the fuzzy approach. By quantifying the factors through the denition of fuzzy membership functions, evaluation of the factors and strategies is made possible and both qualitative and quantitative aspects of the factors are considered. The major approach of the presented algorithm was that in most cases the internal and external factors can not be fully recognized as positive or negative, because their impact on the organization could be observed within a wide spread which may include both positive and negative effects. On the other hand, the aggregation of internal and external factors which leads to extract a strategy in a usual matrix would depend on the intensity and inuence of the factors in this algorithm. Hence, these fuzzy membership functions and the extracted strategies can be well prioritized and it may be possible to concentrate upon strategies with higher priority in implementation stage. Although this paper has attempted to extract strategies based on fuzzy approach, the fuzzy strategy has not been dened. It seems now that the origin of the strategy (that is the internal and the external factors) and the tools for its extraction (SWOT) can be dened in fuzzy environment, but the crisp strategy is extracted as the result of these factors and then in the next step and future researches the fuzzy strategy should be dened. Perhaps the solution for this problem would be to extract fuzzy strategies and then to defuzzify them during a period of time based on the trend of changes. Different researchers are invited to solve this problem. A future study can be to combine more than 2 fuzzy factors for extracting a single strategy.

[3] [4]

[5]

[6]

[7]

[8] [9] [10]

[11]

[12] [13]

[14]

[15] [16]

[17]

References
[1] G. Buyukozkan and O. Feyzioglu, A fuzzy logic based decision making approach for new product development, International Journal of Production Economics 90 (2002), 2745.

[18]